Tooth surface material loss caused by gear wear alters the surface morphology of gears, which impacts their vibration, noise, and remaining lifespan. Although gear wear modeling and prediction have been extensively studied, this paper proposes a novel approach based on the theorem of degradation entropy generation (DEG). A point-by-point calculation method is introduced to determine the degradation coefficient for each measurement point on the tooth profile, accounting for varying working conditions along the tooth profile during the actual meshing process of the gear pair. First, the FZG gear’s bearing capacity is tested. Next, a surface roughness profilometer is employed to in-situ measure the tooth profile after each load stage. The profile deviation curve and the amount of profile wear following each load stage are obtained by processing the measured profile morphology data. Then, the pitting safety factor for each point on the tooth surface is calculated according to the ISO 6336-22:2018 standard and used to correct the degradation coefficient for that point. Finally, the entropy generation of the system during each load stage of the FZG gear is calculated. The degradation coefficient suggested in the DEG theorem is employed to link gear wear with system entropy generation, realizing gear wear modeling and prediction. The results demonstrate that the gear wear calculation method based on the DEG theorem can accurately predict the evolution of tooth profile surface morphology during the experimental process. This research provides a unified calculation method for surface morphology evolution caused by gear wear during service.